UCLA

Understanding the mechanical behavior of solids at small scales is of significant importance for the development of advanced structural materials with superior strength and ductility, and more specifically for the manufacturing of newer class of nano and micro electro-mechanical devices. Every year, new devices are being launched with on-board functional components that are smaller than their previous iterations. The sensors and actuators installed in smartphones to space satellites have now shrunk down to length scales that are below 1 micrometer. Hence, there is a need to characterize the properties of materials at such small scales before they can be employed in the electro-mechanical devices. Majority of materials behave very differently at the nano and micro scales compared to their bulk counterparts. This is primarily due to a reduction in the size and density of the defects. For instance, in crystalline solids, like FCC and BCC metals, the phenomenon of “smaller is stronger” is universally observed, where reducing the characteristic length scale below the micron level results in a significant increase in the material yield strength. This dissertation focuses on uncovering other significant inelastic mechanisms in crystalline as well as amorphous solids, under a wide range of loading conditions. Inspired from previous simulations and experimental works at small scale, a hypothesis is developed whether “reducing domain size to microscale could be a proxy for increasing the temperature”. Since, an increase in situ temperature generally results in reduction in the effective yield strength, the material then should exhibit enhanced ductility at small scales. While the literature has provided evidence of such behavior at lower strain rates in selected materials, the question remains whether such a hypothesis would hold even under dynamic ultrahigh strain rate loading conditions. The aim of this dissertation is to explore the above hypothesis for the entire range of loading strain rates. In addition, this thesis will explore if the above hypothesis is universally valid by choosing two very different material types. The first one is single crystal BCC tungsten (W), which is intrinsically brittle at room temperature in its macroscale form. The second class of material studied in this thesis was the amorphous silica glasses which are among the most abundant materials present on earth. For both materials we demonstrated that reducing the sample size to nano and micro scales resulted in enhanced ductility. For silica glasses, we also observed that the stress required to transform its amorphous phase to its crystalline form was significantly reduced. In fact such a transformation is credited with enhancement of material ductility. Both these phenomena in W and silica glasses are well known to occur at macroscale samples at low stresses only when temperatures are sufficiently high.

Systematic compression and bending experiments were carried out on <100> oriented nano and micro sized pillars and notched-beam specimens of W, respectively. Both types of structures were fabricated via focused ion beam (FIB) milling method and tested using a picoindenter equipped inside a scanning electron microscope (SEM). First, compression experiments at lower strain rates of $ 10^{-3} $ to $ 10^{-1} $ s$^{-1} $ were performed on pillars of diameters between 100 nm and 2 \textmugreek m, with an aspect ratio (AR) equal to 3. The analyses of their stress strain curves revealed that there is a critical size of 500 nm (pillar diameter) below which the plastic flow characteristics, such as the flow stress and strain bursts statistics, become strain rate independent. Second, bending tests on notched cantilever beams with uncracked ligament lengths of 500 nm and 1 \textmugreek m were performed. The crack propagation was visually recorded, along with the force displacement curves which showed slow and stable crack growth. Even with 2 to 3 times higher yield strength, submicron sized beams exhibited similar crack resistance behavior as their macroscale counterparts. Thus, the combination of compression and bending experiments conclusively demonstrate that W can behave in a ductile manner when its characteristic length scale is reduced below 1 \textmugreek m. The key mechanisms behind this phenomenon were found to be the enhanced mobility of screw dislocations, driven by extremely high stress levels. Additionally, higher density of mixed dislocations was observed whose mobility is controlled by phonon drag and less dependent on thermal activation. The experimental data was consistent with the above mechanism that was uncovered by previous DDD simulations.

For exploring the size effect in silica glasses, a series of quasi-static and shock compression experiments were conducted on fused silica (FS, 100\% SiO$_2$) and soda lime glass (SLG, 70\% SiO$_2$) samples. The sample types varied from flat plates to nano and micro pillar geometries, which in turn were fabricated using a combination of FIB and photolithography techniques. A laser generated microflyer plate impact setup was developed for shock loading the glass samples. This setup is capable of launching Al discs of 0.8 to 1.5 mm diameter, and 25 to 50 \textmugreek m thickness, up to speeds ranging between 0.5 and 3.8 km/s. The microflyer generation and its launching was accomplished by focusing a high energy Nd:YAG laser pulse of 8 ns nominal duration with a special flat-head profile. In macroscale geometry, FS is known to undergo phase transition to crystalline Stishovite at shock stress of 34 GPa. However, no such threshold is known for the SLG material. Therefore, we started with shock compression experiments on SLG plates and analyzed the post-shocked samples using transmission electron microscopy (TEM). A stress threshold value of 7 GPa for Stishovite nucleation was found. This is the first report of Stishovite nucleation stress in the SLG material. Diffusion of cationic impurities under localized shear and friction heating is attributed to such a low crystallization stress in SLG.

To establish the effect of domain size on the polymorphic activity, quasistatic compression of FS and SLG nanopillars (500 nm diameter with an aspect ratio equal to 2) was conducted inside an SEM, using the same setup that was used for the W study. Both materials showed unprecedented plastic flow with strains reaching above 50\% and a complete absence of brittle failure. Furthermore, TEM analysis of deformed SLG pillars revealed 4 nm regions of Stishovite crystals. The corresponding stress for this transformation was estimated to be only 4.2 GPa, which is 40\% of the Stishovite nucleation stress of 7 GPa observed in macroscale samples. The diffusion of cationic impurities under very high shear stresses in the SLG pillars is explained as the key mechanism for such a low Stishovite nucleation stress. Interestingly, no Stishovite crystals were observed in any of the FS TEM samples, regardless of the pillar size, where the maximum stress was limited to only 7.2 GPa. Thus, it appears that Stishovite nucleation in FS pillars requires much higher stresses. Unfortunately, such an enhancement of the stress would have required further reduction in pillar size which was not possible due to constraints of the manufacturing procedure and equipment.

However, to overcome the above limitations, the stress levels in the FS pillars was enhanced via shock loading by impacting them with microflyer plates launched at speeds ranging from 0.5-3.8 km/s. For this study, FS pillars of 3 - 15 \textmugreek m in diameter and with an aspect ratio of 1 were fabricated by using the photolithography technique. X-ray diffraction analyses of the shocked samples showed crystalline peaks corresponding to Stishovite nucleation. The corresponding stress was estimated to be 15.24 GPa in the smallest pillars of 3 \textmugreek m diameter. This stress is almost half of 34 GPa that has been widely accepted as the Stishovite nucleation stress for bulk FS. These observations conclusively demonstrate that reducing the domain size to nano and micro scales results in a significant reduction in the threshold stress for the onset of polymorphic activity in silica glasses.